amoeba commented on PR #33918:
URL: https://github.com/apache/arrow/pull/33918#issuecomment-1411355499
Thanks @paleolimbot. Here are two examples:
**Example One**
```r
s3_bucket("my_bucket", endpoint_override = "http://example.com";)
```
- **Current behavior:** Attempts to resolve the bucket's region and does the
wrong thing no matter what
- If a bucket of the name "my_bucket" exists in any AWS region,
erroneously returns that bucket instead of the user's desired bucket
- If a bucket of the name "my_bucket" _does not_ exist in any AWS region,
errors by mistake
- **Desired behavior:** Just trusts the user and sets up the S3FileSystem as
requested with endpoint overriden. Always succeeds.
**Example Two**
```sh
$ cat ~/.aws/config
[default]
region=us-west-2
$ Rscript -e 'arrow::s3_bucket("voltrondata-labs-datasets")$region'
[1] "us-east-2"
# or
$ AWS_REGION="us-west-2" Rscript -e
'arrow::s3_bucket("voltrondata-labs-datasets")$region'
[1] "us-east-2"
```
- **Current behavior:** Successfully resolve the bucket's region but ignores
the user's preference for `us-west-2`. This is only subjective wrong.
- **Desired behavior:** Skips region resolution and sets the region to
`us-west-2`
I think the choice to simultaneously support buckets passed as URIs and also
to construct bucket information via named arguments is what's in conflict here
and eliminating one for the other would move this forward.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
